In population receiving an ambulatory oral treatment, one of the major information, the time at which the patient takes (or not) is medication is unknown. Recent advances (electronic systems recording time of opening and closing of the pill box by the outpatient) in the measure of compliance (CO) with prescribed drug regimens authorities a better knowledge of the exact time of drug administration. Despite these advances, recent and limited statistical methodology has been developed, using only the global compliance defined as the fraction of the total prescribed dose over a time period. No attempt has been made to use the time erratic component of CO which is now available. Population pharmacokinetic-pharmacodynamic (PK-PD) models, that have been developed in order to estimate the statistical distribution of PK-PD parameters (as half-lives or concentration of drug producing half of the maximum response), from sparse data, represent a natural frame for using this information. This proposal is designed to define a methodology for the use of dynamic, time-dependent influence of compliance (TCO) on population pharmacokinetic-pharmacodynamic and/or clinical response (CR), among an outpatient population with prescribed drug regimens. To achieve this goal, the project will be divided in four steps: (i) study of individual profiles of compliance in order to identify a possible population structure of TCO; (ii) use of this structure of TCO for simulating population PK studies, using Monte-Carlo methods, and estimation of the part of inter and intra individual variability that can be attributed to imperfect simulated TCO; (iii) same as (ii) for pharmacodynamic response; (iiii) investigation of possible discrepancies between dose-clinical response and compliance-response functions.